Image degradation refers to the loss of quality or introduction of artifacts in an image, often due to processes like resizing, compression, noise, blur, or distortion. For instance, a high-quality image can degrade when compressed heavily into JPEG (blocky artifacts), or when captured in low light (noise and poor contrast), or if the camera focus is off (blur). In research, certain models aim to be robust to image degradation (like being invariant to noise or blur), or to undo degradation via image restoration techniques. Understanding common degradation processes is important for tasks like super-resolution (undoing low-res degradation) or denoising. Also, data augmentation sometimes intentionally degrades images to make a model resilient (e.g., adding Gaussian noise during training). The term highlights any departure from an image’s ideal quality.